Summary Rather than deficiency models, a concentration-function approach is a more rational approach to determine vitamin intake recommendations for humans. Vitamin C is used as a model vitamin for this work. A concentration-function approach is critically dependent on understanding vitamin C biochemistry and molecular biology; vitamin C clinical pharmacokinetics; and vitamin C physiology. To study concentration-function and physiology relationships in humans for vitamin C, it is necessary to choose clinical samples that can be obtained easily; that contain the vitamin; and where changes in vitamin C concentration may affect physiology and/or cell function. For these reasons, we focused on human red blood cells. As a prerequisite, we developed a new method to measure vitamin C in human red blood cells. This method is the foundation for exploring new functions of vitamin C in red blood cells in health and disease. Humans, unlike most animals, cannot synthesize vitamin C and instead must obtain it from diet. Healthy humans who eat at least 5 servings of fruits and vegetables daily will obtain 200 mg or more of vitamin C. This will produce steady-state fasting plasma concentrations of 70 -80 umoles per liter. Ingestion of more vitamin C from foods will not produce higher concentrations. Even if a vitamin C supplement is taken, plasma concentrations will only rise transiently. All tissues, except red cells, accumulate vitamin C against its plasma concentration. However, once plasma concentrations reach 50 to 60 umoles per liter, tissue concentrations are saturated and do not rise further. Thus, vitamin C concentrations are tightly controlled. Why tight control occurs is unknown. We postulated that one explanation is that tight control of plasma concentrations could facilitate paracrine, or local, actions of the vitamin, if such concentrations were higher. To test this hypothesis, we investigated whether paracrine secretion of vitamin C occurs from adrenal glands in humans. As part of the diagnostic evaluation of patients with hyperaldosteronism, we administered adrenocorticotrophic hormone intravenously to 26 patients. Under radiographic guidance, catheters were placed in both adrenal veins, and adrenal and peripheral blood samples were taken after adrenocorticotrophic hormone was administered. Vitamin C and cortisol concentrations were measured in 47 adrenal veins and 26 peripheral veins. Following adrenocorticotrophic hormone, adrenal vein vitamin C concentrations increased in all cases. The mean peak value increased to 176 +/- 71 umoles per liter, reached between 1 and 4 minutes after adrenocorticotrophic hormone, while peripheral values were unchanged at 35 +/- 15 umoles per liter. Adrenal vein vitamin C concentrations declined nearly to baseline 15 minutes after adrenocorticotrophic hormone. In all cases, adrenal vein vitamin C release preceded cortisol release. The findings were statistically significant with p values for differences < 0.0001. These data are the first description in any species of simultaneous adrenal vein peripheral vitamin C concentrations before and after ACTH stimulation; are the first demonstration that the function of secreted vitamin C must be local rather than systemic; and are the first example of hormone mediated secretion of any vitamin in humans. The data indicate that adrenal vitamin C release in humans is an integral part of the stress response. The data validate the hypothesis that one reason for tight control of vitamin C concentrations is to facilitate a paracrine function. The function of secreted vitamin C in adrenals is unknown, and will be the subject of future research.